1) Write endParaRPr so that PowerPoint display them properly.
2) Original design forbid properites with default value to be exprted, now fixed.
3) Automatic colors are written as white or black based on
   whether background is dark.

Note that tdf#77881,tdf#80520,tdf#89525 depend on this.

Change-Id: I255c16f35149b738be2daf2800b1c90389f2c7cf
Reviewed-on: https://gerrit.libreoffice.org/17472Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>
Reviewed-on: https://gerrit.libreoffice.org/19862Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>
(cherry picked from commit 0b5e3885)

various uninitialized variables in CppunitTest_sd_filters_test hang-8.ppt

Change-Id: Id0b9f146a7ab8b5bb2b36a5c2a39d65ee52e122d
(cherry picked from commit 78407eea)
Reviewed-on: https://gerrit.libreoffice.org/20157Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit d130afe6)

(cherry picked from commit b1e5983f)

Change-Id: I493b79e70fb8003ce215e602f9b4ce92afa50150
Reviewed-on: https://gerrit.libreoffice.org/20156Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>
(cherry picked from commit 2dab392c)

With simple transliteration, TextSearch::searchForward used to use
whole string to perform the search, then started to create substring
to search. But it left the precautions from
commit c00601da by Eike Rathke:
searching for $ may actually return a result set pointing behind the
search string which it does with the ICU regex engine.
The precaution made it to skip reverse mapping if index was 0.

Commit 9aae521b by Michael Stahl
didn't consider the case when nStop is 0, and startPos > 0. Then it
tried to get offset[-1].

Anyway, using value of startPos in those conditions seems illogical.

Removed those precautions (and made assertions for that).
Fixed handling zero indexes.

Reviewed-on: https://gerrit.libreoffice.org/19840Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 4cf1d290)

remove comment that makes no sense, tdf#94810 follow-up

nStart should not become <0, not even when searching for $ end.

(cherry picked from commit 67ab2ce3)

similar to searchForward() use the correct offsets, tdf#94810 related

(cherry picked from commit ce91f3c1)

Change-Id: I2066abc51fff7fb7323bc7f6198bdea06439d4f3
79586471f1a7908372fa3ee044a6f66499b1790f
cb5a45acaad9029e0998594de8b3dc2e9211866e
Reviewed-on: https://gerrit.libreoffice.org/20136Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mike Kaganski <mikekaganski@hotmail.com>

(cherry picked from commit 028d7d84)

On one hand, neither our binary DOC import, nor Word maps the "TOC
Heading" style to something special, and that's how the DOCX import
added that property to some paragraphs in the document, moving the
as-char picture from the first to the second page.

OTOH, the DOCX export filter has a lcl_guessQFormat() function that
explicitly assumes that such a style name exists in Writer document
models, so again it doesn't make sense to handle this style name with
special care.

Change-Id: I3af548930f9683695fc3ad56b486e013f107d61a
(cherry picked from commit 70f3c595)
Reviewed-on: https://gerrit.libreoffice.org/20037Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit 0750162d)

Change-Id: I479a3083f56631135a7f6d94e2864583a9b30a30
(cherry picked from commit 660e37ad)
Reviewed-on: https://gerrit.libreoffice.org/20101Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit afa3f711)

Change-Id: I7896d1d064cd92fd45a6b32ac64f534f68213169
Reviewed-on: https://gerrit.libreoffice.org/20064Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 96ff6372)
Reviewed-on: https://gerrit.libreoffice.org/20126
(cherry picked from commit bad71e69)

It seems that number recognition in tables are not working properly
enter 10-10-10 and it should be converted to a date but it is not.
I tracked it down to the fix of bug fdo#32082. It looks like bSetNumFmt
was changed to false by mistake. Since then it has changed name to
bSetNumFormat. From what I can tell fdo#32082 still works after this
patch, but I might have missed some nuance of that bug report.

Added two tests, one for the bug mentioned above and one to check
that number recognition is working. At least with a simple date.

Reviewed-on: https://gerrit.libreoffice.org/19563Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
(cherry picked from commit aa334d55)

Change-Id: Id58849a223eb602054c66c7379cd56a68a93dea2
Reviewed-on: https://gerrit.libreoffice.org/20082Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit d65ba563)

this fixes a11y issue, where widgets inside VclScrolledWindow weren't
accessible by keyboard as Tab, Alt-Tab events respectively, never
reacheed those
Change-Id: Ic73e24f8036b0f6129c3a4f5efb160f9089cb0b8
Reviewed-on: https://gerrit.libreoffice.org/19982Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

(cherry picked from commit 377381e2)

ODF export in xmloff does the same

Reviewed-on: https://gerrit.libreoffice.org/20051Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Andras Timar <andras.timar@collabora.com>
(cherry picked from commit 11b7a2c8)

Change-Id: I78ddbef3ac96f10f334c45a467dafb95bb9e3453
Reviewed-on: https://gerrit.libreoffice.org/20066Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit 7cb41280)

- drop libxml2-freebsd.patch.1 (upstream libtool 2.4.6 does the same)
- drop libxml2-vc15.patch (fixed upstream)

Change-Id: Ia2f194f39efebd3d2ea924d23a5543ac53e93116
Reviewed-on: https://gerrit.libreoffice.org/20084Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
(cherry picked from commit 391bcf34)
Reviewed-on: https://gerrit.libreoffice.org/20095Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit c8c41401)

(cherry picked from commit 45d1abef)

Change-Id: I733cd21f6d8c7ea1e01f594d1483ad9c2043c188
Reviewed-on: https://gerrit.libreoffice.org/20094Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit 5211c705)

Reviewed-on: https://gerrit.libreoffice.org/18792Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 5ba0f79e)

Backported to 5-0 replacing the interim forward compatibility of
4a334351

Change-Id: Ic053342fde436a7053c15e32683e09b9e91f5308
Signed-off-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 656667ae)

Change-Id: I411a7e26f536a1e4924f566a2ecba097fc134db2
(cherry picked from commit 24271e34)
(cherry picked from commit 10b190ee)

it looks like draw:name values are supposed to be unique in ODF, even
if it's not spelled out explicitly, since it exists so the frame can be
referenced, which sort of implies that it has to be unique, so a
document where the values aren't unique can be considered invalid

    19.197.10 <draw:frame>
    The draw:name attribute specifies a name by which a <draw:frame>
element can be referenced.

So reject duplicate frames but limit this to document generated with
4.4.X which we believe is the version that created these bogus docs

(cherry picked from commit de0432a9)

Change-Id: I83f6d72fd969f667f0a8c2c85d2ffeeed672387a
Reviewed-on: https://gerrit.libreoffice.org/20054Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
(cherry picked from commit df58e602)

XAccessibleEventBroadcaster::addAccessibleEventListener test calls the
fireEvent() and the problem is that the event is triggered not on the
print preview itself, but some toolbar button.  This then causes an
accessible event to be broadcast from the print preview too, but only
after doAccessibleAction() returns, some time later from VCL main-loop.

If the test checks the flag in the listener before the main thread sends
the event, the test fails; try to fix that with
XToolkit::processEventsToIdle().

(cherry picked from commit 1d87c245)

qadevOOo: sadly XToolkit::processEventsToIdle doesn't fix...

... the ScAccessiblePreviewTable test, somehow the event still doesn't
fire in time, let's add an arbitrary sleep too.

(cherry picked from commit c94cf0cf)

Change-Id: Ia52bdf99cdc349ffb0a03536fba271b792caae0f
Reviewed-on: https://gerrit.libreoffice.org/20018Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit e3ee6df3)

Modify styles.xml to not use japanese language specification in style
for date

Change-Id: I0e418c06d5c8d6b04eeaa3e6f6c33c194ee2ef3f
Reviewed-on: https://gerrit.libreoffice.org/20063Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
Tested-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
(cherry picked from commit ec544c4b)
Reviewed-on: https://gerrit.libreoffice.org/20071Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
Tested-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
(cherry picked from commit 7ff72979)

Regression of b5cd11b4

It was always wrong to propagate ForceArray already if a function had a
ForceArray parameter *somewhere*, we need to do this per parameter
instead.

(cherry picked from commit 49257e1d)

Conflicts:
	formula/source/core/api/FormulaCompiler.cxx

prevent ForceArray propagation on the same token, tdf#95670 follow-up

This may happen if the last RPN token is put and the function has a
ForceArray parameter but now again would be wrong if not all parameters
are ForceArray.

(cherry picked from commit b31b17e5)

890fb6b5b88337033cfcf2e8189371ee39461205

-Werror,-Winconsistent-missing-override

(cherry picked from commit 6edc492e)
Backported to SAL_OVERRIDE

3bb2764b625d44f6e0cecbdde3363440faef63b5

Change-Id: If188d45366279d9a7bf641edc7e4dd7095d6d035
Reviewed-on: https://gerrit.libreoffice.org/19993Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit 29adff38)

Change-Id: I2ae5a06f9ffe29de0b1cc5875615010da75f4ae9
(cherry picked from commit d1225d60)
Reviewed-on: https://gerrit.libreoffice.org/20020Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit d9b6f836)

Change-Id: I1d962ad637f19b02855616edebcedbad719689c5
(cherry picked from commit ee655627)
Reviewed-on: https://gerrit.libreoffice.org/20053Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit 22f51321)

... file format by default if the requested filter does not exist.

Much better to write an error message and do nothing.

Change-Id: Ie5404772e7aae5751126bd4c2784b58177804448
(cherry picked from commit c994ce8a)
Reviewed-on: https://gerrit.libreoffice.org/20007Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit de7a50d6)

This fixes a buffer overflow writing over the end of pCaretXArray,
which can happen e.g. when drawing mnemonics in the password dialog.
Based on a similar calculation of nCurrIdx found in
GenericSalLayout::GetCaretPositions().

Change-Id: I7d723cf8cfaeb66f340c7d9ea5b3bc728c6d6209
Reviewed-on: https://gerrit.libreoffice.org/19385Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit f630b6d5)
Reviewed-on: https://gerrit.libreoffice.org/20031Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit e401cb5c)

Change-Id: I7c732d94346be8f37fba1a32655ba224c74e0235
Reviewed-on: https://gerrit.libreoffice.org/20017Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>
(cherry picked from commit aaf3bc89)

substreams require a Model and a DrawPage

Change-Id: I701cf6e2283f84f18b9a0c06a4ce0aa8be934bf3
Reviewed-on: https://gerrit.libreoffice.org/19951Reviewed-by: Oliver Specht <oliver.specht@cib.de>
Tested-by: Oliver Specht <oliver.specht@cib.de>
(cherry picked from commit f47bd056)
Reviewed-on: https://gerrit.libreoffice.org/19954Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit aac3a50a)

so LibreOffice thinks its window is smaller than what gtk3 has allocated
for it.

For gtk3 (like firefox does) split size and position handling, leave
position/move handling to the configureEvent, but listen to sizeAllocate
and use that for the size handling.

Leave gtk2 as it always was

(cherry picked from commit 6adfbe62)

Change-Id: Ic52d6971595741ed658247b651e9e16c2ef9ed0b
Reviewed-on: https://gerrit.libreoffice.org/19924Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit 717c4a8c)

Change-Id: I59a9401c0bdc2d641ea54498af0eb1e589ab12f3
(cherry picked from commit a3223557)
(cherry picked from commit 38e24f1d)
Reviewed-on: https://gerrit.libreoffice.org/19942Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
Tested-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
(cherry picked from commit 1bdffdee)

The following scenario is how it breaks:
1. user has ExcelA1 as her formula syntax setting, CONV_UNSPECIFIED
(that means "same as formula syntax") as her string ref syntax setting
2. she saves the document, it will now contain CONV_UNSPECIFIED value
3. someone else with CalcA1 formula syntax setting opens the document
... since it contains CONV_UNSPECIFIED "same as formula syntax" value,
it will use his CalcA1 formula syntax value to evaluate INDIRECT func
=> #REF!

Avoid this by reading formula syntax/grammar value, mapping it to
matching string ref syntax and saving that instead of CONV_UNSPECIFIED

Change-Id: Ide62d81e6b70c2e9f2ff79648935788107475778
Reviewed-on: https://gerrit.libreoffice.org/19610Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Reviewed-on: https://gerrit.libreoffice.org/19983Tested-by: Jenkins <ci@libreoffice.org>
(cherry picked from commit c3293e0e)

Change-Id: I32f0de20864cae267ebd32516f20baf8e38f910b

Zoom level of SwEditWin is kept in sync with the client, so that the
pixel-based comment widgets can be positioned correctly. But that does
not mean in general the SwEditWin map mode should not be disabled: so
that we don't have to tweak the map mode for each and every
postMouseEvent() call and still be able to send them using logic
coordinates.

Change-Id: I6f686b93d2509d52fdd34e84a502cf04e1ce6e59
(cherry picked from commit de2a6897)

Without this, vcl::Window::ImplTrackTimerHdl() will be called on a
deleted vcl::Window.

Can be reproduced with a comment having a scrollbar in a LOK client,
then clicking on the down button of the scrollbar a number of times ->
crash on exit.

Change-Id: I5d67f96e8baa199f65ec5cf39cb5d39c8162ff33
(cherry picked from commit 7c654ee9)

With this, if a comment has a vertical scrollbar, then not only the
buttons of the scrollbar can be clicked on, but also the slider of the
scrollbar can be dragged.

(cherry picked from commit d562df03)

Conflicts:
	sw/source/uibase/docvw/SidebarWin.cxx

Change-Id: I2e39e18bf60c42a878bb8bfd808f1d47be27eecb

This is similar to the mouse button down handling. When the map mode is
disabled and the map mode is in twips, then in general it's possible to
send mouse coordinates in twips. The scrollbar is usually in pixels, so
add extra code to still make this possible.

Change-Id: I0c7e404ecd7ac839e000266e396683bb7d15c505
(cherry picked from commit d7acb44b)

If a comment had a scrollbar, and the user clicked on the down arrow of
it, then the button remained in the "pushed" state, as the scrollbar
invalidations were not routed to the LOK client.

With this, the button gets back to its non-pushed state after the mouse
button is released.

(cherry picked from commit 55040ea1)

Change-Id: Ie4ba5d0ec07229b0cfc08532e8e91ae25f7a4c9e

(cherry picked from commit 527190f0)

Conflicts:
	sw/source/uibase/docvw/SidebarWin.cxx

Change-Id: Ib1c334825a6629224fe0c8fba564656d53e67410

(cherry picked from commit 43976548)

Conflicts:
	vcl/source/control/scrbar.cxx

Change-Id: Ieaedaa525d613f8ba55f336c613da163a09f7a33

With this, if a comment has enough content that it gets a vertical
scrollbar, then tiled rendering output looks OK, even with non-100%
zoom.

Change-Id: I699aadc11b6c34fb0791e70705719fd61169d972
(cherry picked from commit 30b511ae)

This one is special, as normally its map mode is in pixels, but we need
all sub-widgets to work in twips when tiled rendering.

With this, the scrollbar widget (both the buttons and the
button/background area of the scrollbar itself) is painted at the
correct location when Writer comments have enough content so the
scrollbar is visible.

Change-Id: I4ee9ef8618974b965339078d2262364ec19732ef
(cherry picked from commit f411ef1b)

These member functions take logic units, and the only case when omitting
PixelToLogic() all the time works is when both the widget itself and
render context has a pixel map unit.

Add missing PixelToLogic() calls where the input is in pixels, so in
case the render context has some logic units, the output is still
painted at the correct location.

This is needed for e.g. tiled rendering of the scrollbar of Writer
comment widgets.

Change-Id: Ideb75625b1f6349c5f18342e6f438fd5d7c31dcd
(cherry picked from commit f2081039)

Checking for a bool is certainly cheaper than a chain of member function
calls.

Change-Id: I9520194e58e75264f3fcbb413a618e3995f59a4e
(cherry picked from commit 5eb7759d)

Instead of invalidating the whole area, invalidate the sub-widget when
the whole area of the sub-widget would be invalidated.

With this, a test comment with enough comment to have a scrollbar
results in 3 paintTile() calls instead of 11 ones (70% save).

(cherry picked from commit 9ece5c87)

Conflicts:
	sw/source/uibase/docvw/SidebarTxtControl.cxx

Change-Id: I2fe317549eefac9a63aaf50f5a9a242e15c4dc86